Texas's Power Pause Puts a Price on Grid Access
A 474-gigawatt queue and a new state audit turn data-center land from a power-access play into a power-allocation bet.
Texas has been the place data center developers go for easier access to power, and Bisnow reports the state is the nation's fastest-growing data center hotbed, on track to surpass Virginia as the industry's largest market by 2030 on the strength of that access and a business-friendly political climate. That climate now carries a new layer of review: Gov. Greg Abbott, a Republican who previously supported data center expansion, has ordered a halt to all data center connections until state agencies complete an audit of each project's planned power and water consumption, ownership, and related details.
The pause arrives ahead of midterm elections in which opposition to data centers has become a central issue for candidates in both parties, according to Bisnow, and it puts every project awaiting grid access in a state of limbo: not only when a connection will arrive, but whether it will arrive at all. That uncertainty split the industry at Bisnow's DICE: South event in Arlington earlier this month.
At the event, some Texas developers called the edict a mistake, arguing it is a politically motivated measure that threatens to stop the state's digital infrastructure boom, while others, including leaders of large data center firms, said the pause is necessary because the audits will decide which projects receive power first and winnow out those that should not be in the queue. Skybox Datacenters Chief Development Officer Haynes Strader said the state needs a firmer grasp on who is getting interconnected so responsible players proceed and the right narrative holds.
Power becomes the underwrite
For private real estate capital, the policy fight matters less than what the pause does to valuations: Texas data-center land was being priced as if power access were a predictable entitlement, and the freeze exposes that assumption. Access is now a state allocation decision, and the connection date is a variable that can stretch development timelines and reset carry costs; a project with land but no grid connection is a different asset today than it was before the order.
Connection dates dominate the math because the Abbott administration says the Electric Reliability Council of Texas is weighing more than 474 gigawatts of requested power capacity from large projects—more than five times the record demand on the state grid and the equivalent of serving more than 415 million new homes—with data centers accounting for 90% of those requests. No credible reading of that pipeline treats it as a list the grid will clear in any reasonable period.
The audit list is the new underwriting document: each project will be judged on its planned power draw, water use, and ownership, so sponsors with fully documented plans gain an edge over sponsors who filed speculative requests. As this publication has argued, data-center debt is becoming its own asset class, and the construction-cost squeeze will separate sponsors who can source power from those who cannot—the Texas pause is the policy version of that same test.
For investors, the pause is the first real test of the idea that any Texas site can be monetized as a data center because demand is inexhaustible. The queue may express demand, but demand does not equal deliverable product, and a sponsor with a secured grid connection is holding a project while a sponsor with a place in line is carrying a cost center. Development underwriting will now need a line item for the probability and cost of waiting out a state review; that cost will not be zero.
Sponsors who can document their load today have options, while sponsors whose thesis was a cheap option on Texas land are now carrying a decaying one. The state's review is effectively rationing access, and every week of review raises the cost of capital for projects without a connection—for institutional investors, the connection itself is the valuation.
The pause does not end the Texas data-center boom, and it likely should not, but it does reset the risk model. The projects that survive will be the ones able to demonstrate real load, and the capital behind them will see a market with fewer, better-documented competitors. The first tranche of approved connections will be the moment Texas data-center pricing reveals its new level; investors watching the queue, not just the headlines, will be positioned for that repricing.